lecture temperature transducers 2
TRANSCRIPT
-
8/13/2019 Lecture Temperature Transducers 2
1/15
1
Temperature Sensors
What is the best sensor to measure temperature?
There is no such thing depends on the temperature range,
the accuracy needed, the environment, the cost, etc.
We have looked at RTDs, which have their electrical
resistance vary with temperature. We then studied how to
measure resistance.
NOTE: While we talked about measuring resistance in the
context of thermometry, the tips we used are general, andapply to measuring resistance for other reasons too. We willrevisit this later today and in future lectures.
Thermistors
Solid-state semiconductor device
Positive temperature coefficient (PTC) devices
increases resistance with temperature increase
Negative temperature coefficient (NTC) devices
decreases resistance with temperature increase
Relationship between R and T is non-linear,
but has a very steep slope
Increases the sensitivity of the device Resistance change of 3%/C
Limits range of operation
-
8/13/2019 Lecture Temperature Transducers 2
2/15
2
Thermistors
Non-linear resistance vs. temperature means:High sensitivity
Reproducibility not as good as Pt RTD
Calibration important, and must be done at several
points.
Resistance can be chosen: !to M!
Temperature range small, but can be chosen in
range of interest
Devices can be tiny - therefore fast response time. Muchcheaper than Pt RTDs.
Chemical stability not as good long term drift due to
thermal cycling.
Thermistor vs RTD
-
8/13/2019 Lecture Temperature Transducers 2
3/15
3
Thermistors
Thermistor Functional Form
Relationship between R and T given by
the Steinhart-Hart equation
RT- resistance at temperature T
a, b, c, d - constants given by device
manufacturer
T - temperature in K
1
T= a+ bln R
T( ) + c ln RT( )2
+ dln RT( )
3
-
8/13/2019 Lecture Temperature Transducers 2
4/15
4
Higher resistance makes lead wire effects negligible.
Self-heating problem identical to RTDs - best to stilluse a bridge.
Cryogenic SensorCernox from Lakeshore
Each device must be calibrated for
accurate use, but general form is
similar, so a 1-pt calibration often used
-
8/13/2019 Lecture Temperature Transducers 2
5/15
5
Temperature Ranges
Temperature (Kelvin)
Measuring
Temperature
1mK 10mK 100mK 1K 10K 100K 1000K 10000K
4.2K 77K 0oC
100oC
4He
N2
194K
CO2
H2O
Magnetic Thermometry
RuO2Resistance
Pt Resistance
Thermistor
Copper-Constantan
Thermocouples
Iron
Const
antan
Chromel-
Alumel
Tungsten-Rhenium
Infrared
PyrometryMelting Curve
Spectroscopy
Thermocouples
Two dissimilar metals in contact generate a potential
difference between junctions at different temperatures
+V -V
Seebeck EffectRelies on the correlation between the electric potential
and thermal gradient.
The size of the effect, and reproducibility has lead to
established pairs of metals.
-
8/13/2019 Lecture Temperature Transducers 2
6/15
6
Standard Thermocouples
Need to keep a reference
junction at a known
temperature. Tables for
0C (see Omega catalog!).
Electronic compensation
can be used if the
reference junction isnt at
0C.
ThermocouplesWork over a wide range of temperatureNo self-heating
mV output, but reasonable sensitivity ("V/ "T large).
Use the Thermocouple Laws:
Thermal emf unaffected by temperature elsewhere in the
circuit if the two metals are homogeneous
A third homogeneous metal inserted in the circuit doesnt
affect emf [this is the voltmeter].
Cheap! Rugged!
Can be small for fast response.
Good linearity, but not particularly accurate.
-
8/13/2019 Lecture Temperature Transducers 2
7/15
7
Sealed
sheath
Grounded
to sheath
Exposed
Isolated from sheath
+ Wires well protected
- Slow response time (75 sec)
Grounded to sheath
+ Wires well protected
+ Reasonable response time (40 sec)
Exposed
+ Fast response time (2 sec)
- Wires exposed to damage
(physical/chemical)
Thermocouples
-200 - 350 0C55% Cu
45% Ni
100% CuT
0 - 1450 0C100% Pt90% Pt10% Rh
S
0 - 1450 0C100% Pt87% Pt
13% RhR
0 - 1250 0C
95.5% Ni
4.4% Si
84.4% Ni
14.2% Cr
1.4% Si
N
0 - 1250 0C95% Ni
5% other
90% Ni
10% CrK
0 - 750 0C55% Cu
45% Ni
99.5% FeJ
0 - 900 0C55% Cu
45% Ni
90% Ni
10% CrE
870 - 1700 0C93.9% Pt
6.1% Rh
70.4% Pt
29.6% RhB
TemperatureRange
Negative LegMetal
Positive LegMetal
Standard Type
*Modified from Temperature Sensors,
The Watlow Educational Series Book Four
-
8/13/2019 Lecture Temperature Transducers 2
8/15
8
Seebeck Effect
The emf generated by the Seebeck effect can be present inRTDs and thermistors too! This is an unwanted effect that
would look like a different temperature.
rsI
rL
rL
RVrin
rL
rL
emf
Avoid it by reversing current: emf doesnt change, but IRdoes. We can do this with positive and negative DC
currents, or use an AC signal.
Cryogenic AC Resistance
BridgeAt low temperature, the self-heatingeffect is even more pronounced, so
ALL these techniques are used, plus
many more.
P=V2/R
=10-8/103
=10-11W
"R/R=
.1/103=10-4
-
8/13/2019 Lecture Temperature Transducers 2
9/15
9
Junction Semiconductor
Sensors
Silicon transistors configured in a circuit that has
measurable and predictable I-V. Different configurations
possible.
Benefits:
Cheap
Integrated with other silicon circuits (for memory,
communication, amplification, processing, etc.)
Disadvantages
Needs power, not rugged, not standardized, limited
temperature range
AD590 IC Temperature
SensorOperating Principle: Transistors current in saturation dependson
-
8/13/2019 Lecture Temperature Transducers 2
10/15
10
Optical Thermometry
Hot bodies radiateelectromagnetic energy
depending on their
temperature.
M=energy emitted per
unit frequency.
The surface of thematerial matters:
emissivity of a material#
-
8/13/2019 Lecture Temperature Transducers 2
11/15
-
8/13/2019 Lecture Temperature Transducers 2
12/15
12
Pyrometry
For still higher temperatures, the thermal emission will be in the
visible regime.
Measuring the colour of the light,
or more precisely the spectrum,
will allow the temperature to be
known.
One simple way to measure
temperature is to compare two
spectra - Disappearing filamentoptical pyrometer.
Disappearing Filament Optical
Pyrometer
Look at background image while
looking at a tungsten filament. Put
power into the filament until it appears
to glow the same colour as the
background. Using the known
properties of the filament, its
temperature can be found, and the
temperature of the background objectis the same.
-
8/13/2019 Lecture Temperature Transducers 2
13/15
13
Measurement & Uncertainties
Anytime you make a measurement, you should quote anuncertainty on that measurement. This uncertainty tells thereader how confident you are in the number and takes into
account two effects:
Statistical uncertainty:due to the random nature of some
processes, we expect there to be variation in the measured
values. For example: nuclear decay is a perfectly random
event, so counting the number of decays in a given time will
have some variation.
Statistical uncertainties happen whenever randomness is
inherent:
- Nuclear decay
-quantum processes
-sampling of a larger ensemble
Measurement and Uncertainties
Systematic Uncertainties:
Systematic Uncertainty: This is the degree to which a
measured value differs from the 'true' value because of errors
inherent in the measurement. This may be due to an incorrect
scale, a wrong calibration or an erroneous assumption. In
instrumentation, changes in the original calibration of an
instrument over time, or if the instrument is used under
abnormal conditions are major sources of systematic error.Systematic errors are generally not statistical in natureand
may be corrected by measurement of a standard.From textbook
-
8/13/2019 Lecture Temperature Transducers 2
14/15
14
Measurement and Uncertainties
These definitions are rather narrow, and exclude lots of other
errors:
Limited precision of a measurement device
Uncontrolled variation of extraneous variables (eg. biology)
Variation in amplifier output which affects measured values
Difficulty in making a measurement
These fluctuations may or may not be strictly random or
statistical in nature. Often we will use statistics to study these
effects - which may or may not be strictly true. Much of this
course focuses on designing instruments or experiments
which minimize these effects.
Example of Uncertainties
-
8/13/2019 Lecture Temperature Transducers 2
15/15
15
Uncertainties
Physics 252 (or other statistics class) dealt with the random
uncertainties present in all measurements, and harnessed
the power of statistics to deal with them.
- Much of this is reviewed in chapter 1 of
Sayer & Mansingh.
Physics 352 builds on this. How can we use physics,instrumentation and electronics to reduce and understand
these uncertainties? We will still need to understand
statistics, since this is the way these random fluctuationsare characterized, but we will go beyond that to study why
the fluctuations occur
Next few lectures: NOISE